Few indicators of the likelihood of survival have been reported for clinically ill client-owned rabbits admitted to veterinary hospitals. In a study1 involving 322 critically ill rabbits, the mortality rate over a 7-day period for rabbits with hyponatremia (serum sodium concentration, < 129 mEq/L) was 2.3 times that for rabbits with normonatremia. In another study2 involving 907 rabbits, blood glucose concentrations > 20 mmol/L were observed only in rabbits with life-threatening diseases. Furthermore, rabbits with intestinal obstruction had significantly higher blood glucose values than rabbits with gastrointestinal stasis.
Information about the clinical usefulness of BUN concentration in rabbits is limited. One study3 revealed that client-owned rabbits seropositive for Encephalitozoon cuniculi infection had higher BUN concentrations than seronegative rabbits, but both groups had values within reference limits. Another study4 involving 190 anorectic rabbits with BUN values > 27 mg/dL revealed that rabbits with higher BUN values had a higher mortality rate than rabbits with lower values (but still > 27 mg/dL). However, inclusion of only anorectic rabbits with high BUN concentrations limited the generalizability of these findings.
Identification of commonly measured analytes that could be used as prognostic indicators would be helpful in the clinical management of rabbits. The purpose of the study reported here was to evaluate whether BUN concentration in client-owned rabbits at admission to a veterinary referral center was associated with short-term mortality rate. Our hypothesis was that rabbits with high BUN values on admission to a veterinary referral center would have a higher mortality rate over the subsequent 15-day period than rabbits with BUN values within reference limits.
Materials and Methods
Animals
Computerized laboratory records of the Centre Hospitalier Frégis, Arcueil, France from July 1, 2011, to February 29, 2016, were searched to identify rabbits that had BUN concentration measured in a blood sample obtained from a lateral saphenous vein. At this veterinary referral center, BUN concentration measurements were typically performed for clinically ill rabbits and for healthy rabbits as part of a routine health check prior to elective surgery. All rabbits that had a medical record and at least 1 BUN concentration measurement during the period from July 1, 2011, to February 29, 2016, were eligible for inclusion. Rabbits for which follow-up information was available for a period < 15 days were excluded from the study.
Study design and data collection
A retrospective cohort study design was used. Data were collected from the medical record of each rabbit regarding BUN concentration (ie, the exposure variable of interest) and other factors hypothesized to be associated with death due to any cause within 15 days after the first BUN concentration measurement (ie, the outcome variable of interest). These data included breed, sex, age, owner-reported appetite (anorexia or no anorexia), and definitive or presumptive diagnosis. For rabbits with > 1 BUN concentration recorded during the follow-up period, only the first obtained value was included in the study.
If certain information was not available in the medical record or in the log of communications with the owner or referring veterinarian, an investigator (GZ) contacted the owner by the telephone number on record to ascertain survival status of the rabbit 15 days after the first BUN concentration measurement. If the call went unanswered, the owner was contacted 7 days later at a different time of day to a maximum of 3 attempts. When the owner could not be contacted or when the owner was unsure of the rabbit's survival status, the rabbit was excluded from the study.
BUN concentration measurement
Blood samples were collected from rabbits via a lateral saphenous vein with 25-gauge, 0.625-inch needles into 500-μL microtubes containing lithium heparin, without withholding of food beforehand. All BUN concentrations were measured at the same veterinary laboratory by use of an automated analyzer.a
Statistical analysis
The Shapiro-Wilk test was used to analyze continuous data (ie, rabbit age and BUN concentration) for normality of distribution. Normally distributed data are summarized as mean ± SD, and nonnormally distributed data are summarized as median (IQR). Univariate logistic regression was performed to identify factors (age, sex, and owner-reported appetite) associated with outcome (death during the 15-day period following the first BUN concentration measurement). The BUN data were also classified by whether they exceeded or were within standard reference limits (< 23.3 mg/dL)5 and examined for an association with outcome by means of the χ2 or Fisher exact test, as appropriate. Odds ratios and 95% CIs are reported for all analyses to quantify the strength of associations. The RR and associated 95% CI were calculated for mortality rate comparisons between rabbits with a high BUN concentration and those with a concentration within reference limits. Subgroup analyses were also conducted to support the robustness and generalizability of the findings for rabbits with no anorexia, compared with those for rabbits with anorexia.
Variables with values of P < 0.05 identified on univariate analysis as well as sex (considered a likely confounder regardless of significance) were then considered for inclusion in a multivariate logistic regression model. To avoid overfitting the model, no more than 1 predictor variable was included for every 10 events (ie, deaths).6 For this analysis, the assumption for the BUN data was assessed as described elsewhere.7 The data were first grouped by quartiles into categories, a multivariable model was fitted to replace the continuous BUN data with the categorical BUN data, the log odds of the upper 3 quartile categories (relative to the first quartile category) were plotted versus the respective quartile midpoints, the 4 plotted points were connected with straight lines, and the plot was visually inspected for linearity. Because the increase in log odds did not conform to a linear gradient, the quartile-categorized BUN data were used in subsequent analyses.
The Hosmer-Lemeshow test was used to assess the goodness of fit of the final multivariate model. The Nagelkerke R2 value is reported as a measure of the predictive power of that model.
Results
Animals
A total of 269 healthy or clinically ill client-owned rabbits had at least 1 BUN concentration measurement recorded at the veterinary referral center during the study period and were thus eligible for inclusion. Twenty-three (8.6%) of these rabbits (17 males, 4 females, and 2 with no sex recorded) were excluded because they lacked 15 days of follow-up. Therefore, 246 rabbits (139 males, 102 females, and 5 with no recorded sex) met the study criteria and were included in the study. Rabbits were classified in the medical records as lop-eared (n = 69), dwarf (61), lionhead (3), giant (1), and toy (1); no breed was recorded for 111 rabbits. Mean ± SD age at the date of the first BUN concentration measurement was 6.1 ± 2.7 years (range, 0.7 to 15 years).
Diagnoses included dental disease (74/246 [30.1%]), urinary tract disease (43/246 [17.5%]), primary or secondary gastrointestinal stasis (41/246 [16.7%]), neurologic disorder (41/246 [16.7%]), and uterine enlargement (25/246 [10.2%]). Healthy rabbits (13/246 [5.3%]) and clinically ill rabbits with other diagnoses were less common, and many rabbits had > 1 diagnosis. Owner-reported appetite was recorded for 240 rabbits, of which 121 (50.4%) were recorded as having anorexia and 119 (49.6%) as having a typical appetite. Overall, 78 of 246 (31.7%) rabbits died within 15 days after BUN concentration measurement. Rabbit characteristics were summarized by survival status (Table 1).
Results of univariate analyses to identify variables associated with death due to any cause over the 15-day period following initial measurement of BUN concentration in healthy and clinically ill client-owned rabbits at a veterinary referral center.
| Variable | No. of rabbits | Survivors | Nonsurvivors | OR (95% CI) | P value |
|---|---|---|---|---|---|
| Age (y) | 241 | 5.8 ± 2.5 | 6.8 ± 3.1 | 1.14 (1.03–1.26) | 0.01 |
| Sex | |||||
| Female | 102 | 65 (63.7) | 37 (36.3) | 1.46 (0.84–2.52) | 0.18 |
| Male | 139 | 100 (71.9) | 39 (28.1) | Referent | — |
| BUN concentration (mg/dL) | |||||
| ≥ 23.3 | 87 | 50 (57.5) | 37 (42.5) | 2.13 (1.22– 3.71) | 0.007 |
| < 23.3 | 159 | 118 (74.2) | 41 (25.8) | Referent | — |
| Owner-reported appetite | |||||
| Anorexia | 121 | 75 (62.0) | 46 (38.0) | 1.82 (1.05–3.16) | 0.03 |
| Typical | 119 | 89 (74.8) | 30 (25.2) | Referent | — |
Data for age are reported as mean ± SD and for all other variables as number (%) of all rabbits with the indicated factor. For categorical variables (here, sex, BUN concentration, or appetite), a significant OR > 1 indicates an increased probability of the outcome (ie, death within 15 days) relative to the probability in the referent group, and a significant OR < 1 indicates a decreased probability. For continuous variables (here, age), a significant OR reflects the increase or decrease in the probability of the outcome for each 1-unit increase in the variable.
— = Not applicable.
BUN concentration and outcome
Median BUN concentration for all rabbits was 18.7 mg/dL (range, 6.5 to 251.1 mg/dL; IQR, 14.0 to 24.7 mg/dL); 159 of 246 (64.6%) rabbits had an initial BUN concentration within reference limits (< 23.3 mg/dL), and 87 (35.4%) had a high BUN concentration (ie, ≥ 23.3 mg/dL). Median BUN concentration in rabbits that failed to survive for 15 days was 22.2 mg/dL (IQR, 14.9 to 56.9 mg/dL) and for rabbits that survived for that period was 18.2 mg/dL (IQR, 13.9 to 23.3 mg/dL; Figure 1).
Distribution of BUN concentrations in healthy and clinically ill client-owned rabbits with an owner-reported typical appetite (n = 119; A) or anorexia (121; B) that survived (white bars) or failed to survive (striped bars) for at least 15 days after the first BUN concentration measurement.
Citation: Journal of the American Veterinary Medical Association 255, 2; 10.2460/javma.255.2.200
Distribution of BUN concentrations in healthy and clinically ill client-owned rabbits with an owner-reported typical appetite (n = 119; A) or anorexia (121; B) that survived (white bars) or failed to survive (striped bars) for at least 15 days after the first BUN concentration measurement.
Citation: Journal of the American Veterinary Medical Association 255, 2; 10.2460/javma.255.2.200
Distribution of BUN concentrations in healthy and clinically ill client-owned rabbits with an owner-reported typical appetite (n = 119; A) or anorexia (121; B) that survived (white bars) or failed to survive (striped bars) for at least 15 days after the first BUN concentration measurement.
Citation: Journal of the American Veterinary Medical Association 255, 2; 10.2460/javma.255.2.200
Results of univariate analysis indicated that rabbits with a high BUN concentration were more likely to die during the follow-up period than rabbits with a value within reference limits (Table 1). The mortality rate for rabbits with a high BUN concentration was 33% higher than that of rabbits without a high concentration (RR, 1.33; 95% CI, 1.06 to 1.69).
After adjustment for sex, age, and owner-reported appetite, the multivariate logistic regression model indicated that rabbits with BUN values > 24.7 mg/dL had approximately 3 times the odds of nonsurvival over the follow-up period, compared with rabbits with BUN values < 14.00 mg/dL (Table 2). This model fit the data adequately (Hosmer-Lemeshow test, P = 0.32) and explained 14% of the assessed outcome (Nagelkerke R2 = 0.14). The receiver operating characteristic curve based on predicted probabilities from the model had an area under the curve of 0.69 (95% CI, 0.61 to 0.70; Figure 2).
Receiver operating characteristic curve based on the predicted probabilities generated by a multivariate logistic regression model to evaluate whether BUN concentration was associated with death due to any cause over the 15-day period following measurement for the rabbits of Figure 1 (n = 231).
Citation: Journal of the American Veterinary Medical Association 255, 2; 10.2460/javma.255.2.200
Receiver operating characteristic curve based on the predicted probabilities generated by a multivariate logistic regression model to evaluate whether BUN concentration was associated with death due to any cause over the 15-day period following measurement for the rabbits of Figure 1 (n = 231).
Citation: Journal of the American Veterinary Medical Association 255, 2; 10.2460/javma.255.2.200
Receiver operating characteristic curve based on the predicted probabilities generated by a multivariate logistic regression model to evaluate whether BUN concentration was associated with death due to any cause over the 15-day period following measurement for the rabbits of Figure 1 (n = 231).
Citation: Journal of the American Veterinary Medical Association 255, 2; 10.2460/javma.255.2.200
Output of the final multivariate logistic regression model to evaluate whether BUN concentration was associated with death due to any cause over the 15-day period following initial measurement for the rabbits of Table 1 (n = 231).
| Variable | OR (95% CI) | P value |
|---|---|---|
| BUN concentration* (mg/dL) | ||
| < 14.00 | Referent | — |
| 14.00–18.67 | 1.08 (0.45–2.51) | 0.86 |
| 18.67–24.74 | 1.26 (0.52–3.05) | 0.61 |
| > 24.74 | 2.92 (1.29–6.58) | 0.01 |
| Age (y) | 1.17 (1.05–1.31) | 0.005 |
| Sex (male vs female) | 0.61 (0.33–1.10) | 0.10 |
| Anorexia (yes vs no) | 1.45 (0.78–2.67) | 0.24 |
Categories reflect quartile values.
See Table 1 for remainder of key.
Findings of subgroup analysis in which rabbits were considered on the basis of whether they had anorexia suggested a different clinical relevance of high BUN concentration between these groups. Considering only rabbits with anorexia, 52% (30/58) of those with a high BUN concentration survived, whereas 71% (45/63) of those with an unremarkable BUN concentration (ie, < 23.3 mg/dL) survived. Therefore, in anorectic rabbits, a high BUN concentration was associated with an increased mortality rate (RR, 1.69; 95% CI, 1.05 to 2.71; P = 0.03). On the other hand, considering only rabbits with a typical appetite, 69% (18/26) of those with a high BUN concentration survived, compared with 76% (71/93) of those with an unremarkable BUN concentration. Therefore, in rabbits with a typical appetite, no association was identified between BUN concentration and mortality rate (RR, 1.30; 95% CI, 0.66 to 2.58; P = 0.46).
Discussion
Results of the present retrospective cohort study involving 246 client-owned rabbits admitted to a single veterinary referral center suggested that a BUN concentration cutoff of 24.7 mg/dL may be useful for predicting the probability of anorectic rabbits surviving for at least 15 days after BUN concentration measurement. Unexpectedly, in the population evaluated, BUN concentration was not linearly associated with outcome.
Blood urea nitrogen concentration has been identified as a predictor of outcome in both veterinary and human medicine. For example, a high BUN concentration predicts nonsurvival in humans with heart failure or pancreatitis8,9; dogs with intrahepatic portosystemic shunts,10 parvovirus infection,11 or idiopathic immune-mediated hemolytic anemia12; and cats with chronic kidney disease.13 Consistent with our findings, results of studies involving critically ill humans indicate that BUN concentration predicts nonsurvival14 as well as disease severity and outcome.9
Interestingly, BUN concentration (categorized by quartile) in the rabbits of the present study was significantly and strongly associated with mortality rate on multivariate analysis only when > 24.74 mg/dL, compared with the mortality rate for rabbits with concentrations < 14.00 mg/dL. To the authors’ knowledge, no reports exist of the diagnostic sensitivity and specific of BUN concentration measurement for diagnosing kidney disease in rabbits. Therefore, a cutoff of 23.3 mg/dL was used to distinguish high from unremarkable values in the present study because this was the upper reference limit of BUN values obtained in another study5 involving 155 client-owned rabbits of both sexes and various breeds from which blood samples were collected from a lateral saphenous vein and in which BUN concentration was measured by means of an enzymatic technique similar to ours. Because BUN values > 24.7 mg/dL are considered higher than the upper reference limit, it is possible that only BUN values that exceed the upper reference limit predict mortality rate. Although it would need to be confirmed, failure to detect a significant association between BUN values and outcome at every quartile (ie, a dose-dependent association) might have reflected insufficient statistical power attributable to small sample sizes.
The reasons that high BUN concentrations might appear to influence short-term outcomes such as survival over a 15-day period must be considered. One possibility is that diseases associated with a high BUN concentration carry a poor prognosis,15 regardless of whether those diseases are primarily renal or extrarenal in nature. In addition, a high BUN concentration can cause uremia and associated clinical signs and complications, such as gastrointestinal,16 cardiovascular,17 or pulmonary18 abnormalities.
One limitation of the study reported here was that it involved patients of a single veterinary center with a typical caseload comprised of 40% referrals, 25% emergencies, and 35% primary care. The high overall mortality rate of 31.7% suggested that the study sample was composed primarily of critically ill rabbits, and it is possible that the prevalence of high BUN concentrations and the overall mortality rate would differ in another setting. Some data were missing owing to the retrospective nature of the study, thereby leading to a loss of statistical power. However, the number of rabbits lost to follow-up was considered too small to impact the conclusions, and the characteristics of those rabbits were similar to those of the included rabbits. Additionally, a bias in the classification of the outcome might have occurred when the date of the death was not recorded and had to be retrospectively determined. Such bias would have been nondifferential in nature and would not be expected to have substantially influenced the findings. Furthermore, BUN concentration in rabbits is physiologically influenced by several factors such as sex, breed, dietary protein value and quality, withholding of food, and natural diurnal rhythm, with higher concentrations noted in the late evening.19 The precise time of blood sample collection for BUN concentration measurements was not mentioned in the records and may have represented a confounding factor.
The usefulness of BUN concentration as a predictor of short-term mortality rate in rabbits needs to be investigated further. A multicenter prospective study should be conducted, and other potential confounders such as glycemia, natremia, and rectal temperature1,2,20 should be controlled for in future analyses. Blood urea nitrogen concentrations in rabbits should be evaluated with respect to primary disease and temporal changes during hospitalization. The effects of specific treatments on short-term outcomes for rabbits with a high BUN concentration at initial evaluation should also be investigated.
Overall, the present study showed that client-owned rabbits with a high BUN concentration (≥ 23.3 mg/dL) had a higher mortality rate over a 15-day period than did similar rabbits without a high BUN concentration. Moreover, rabbits with BUN concentrations > 24.74 mg/dL had almost 3 times the odds of nonsurvival during that period as did rabbits with BUN concentrations < 14 mg/dL. Therefore, BUN concentration measurement could be valuable for assessing short-term prognosis for client-owned rabbits and assisting clinicians with decision-making during emergency situations.
Acknowledgments
No financial support was received for the study. The authors declare that there were no conflicts of interest.
Presented in abstract form at ExoticsCon, San Antonio, Tex, August 2015.
ABBREVIATIONS
| CI | Confidence interval |
| IQR | Interquartile (25th to 75th percentile) range |
| RR | Relative risk |
Footnotes
Konelab 60 clinical chemistry analyzer, Thermo Fischer Scientific, Waltham, Mass.
References
1. Bonvehi C, Ardiaca M, Barrera S, et al. Prevalence and types of hyponatraemia, its relationship with hyperglycemia and mortality in ill pet rabbits. Vet Rec 2014;174:554.
2. Harcourt-Brown FM, Harcourt-Brown SF. Clinical value of blood glucose measurement in pet rabbits. Vet Rec 2012;170:674.
3. Ewringmann A, Gobel T. Untersuchung blutchemischer Parameter bei Heimtierkaninchen. Kleinterpraxis 1998;43:447–452.
4. Saito K, Hasegawa A. Diseases and outcomes in rabbits with high BUN levels. J Vet Med Sci 2003;65:625–628.
5. Hein J, Hartmann K. Labordiagnostische Referenzbereiche bei Kaninchen. Tierarztl Prax 2003;31:321–328.
6. Peduzzi P, Concato J, Feinstein AR, et al. Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates. J Clin Epidemiol 1995;48:1503–1510.
7. Hosmer DW, Lemeshow S, Sturdivant RX. Applied logistic regression. 3rd ed. Somerset, NJ: John Wiley and Sons Inc, 2013.
8. Cauthen CA, Lipinski MJ, Abbate A, et al. Relation of blood urea nitrogen to long-term mortality in patients with heart failure. Am J Cardiol 2008;101:1643–1647.
9. Wu BU, Johannes RS, Sun X, et al. Early changes in blood urea nitrogen predict mortality in acute pancreatitis. Gastroenterology 2009;137:129–135.
10. Papazoglou LG, Monnet E, Seim HB. Survival and prognostic indicators for dogs with intrahepatic portosystemic shunts: 32 cases (1990–2000). Vet Surg 2002;31:561–570.
11. Bastan I, Kurtdede A, Özen D. Prognostic usefulness of some parameters in dogs with canine parvovirus. Ankara Univ Vet Fak Derg 2013;60:53–58.
12. Piek CJ, Junius G, Dekker A, et al. Idiopathic immune-mediated hemolytic anemia: treatment outcome and prognostic factors in 149 dogs. J Vet Intern Med 2008;22:366–373.
13. King JN, Tasker S, Gunn-Moore DA, et al. Prognostic factors in cats with chronic kidney disease. J Vet Intern Med 2007;21:906–916.
14. Beier K, Eppanapally S, Bazick H, et al. Elevation of BUN is predictive of long-term mortality in critically ill patients independent of “normal” creatinine. Crit Care Med 2011;39:305–313.
15. Waikar SS, Bonventre JV. Can we rely on blood urea nitrogen as a biomarker to determine whether to initiate dialysis? Clin J Am Soc Nephrol 2006;1:903–904.
16. Crivellenti LZ, Borin-Crivellenti S, Fertal KL, et al. Occult gastrointestinal bleeding is a common finding in dogs with chronic kidney disease. Vet Clin Pathol 2017;46:132–137.
17. Bijsmans ES, Jepson RE, Chang YM, et al. Changes in systolic blood pressure over time in healthy cats and cats with chronic kidney disease. J Vet Intern Med 2015;29:855–861.
18. Le Boedec K, Heng HG, Snyder PW, et al. Pulmonary abnormalities in dogs with renal 400 azotemia. J Vet Intern Med 2012;26:1099–1106.
19. Benson KG, Paul-Murphy J. Clinical pathology of the domestic rabbit acquisition and interpretation of samples. Vet Clin North Am Exot Anim Pract 1999;2:539–551.
20. Di Girolamo N, Toth G, Selleri P. Prognostic value of rectal temperature at hospital admission in client-owned rabbits. J Am Vet Med Assoc 2016;248:288–297.
